Social relationships have been reliably related to lower cardiovascular morbidity and mortality. Most of this research has been interpreted as reflecting the beneficial influences of social support. However, the joint contribution of positive and negative aspects of relationships in the context of support processes has hot been adequately considered. In addition, little is known of the mechanisms linking relationships to cardiovascular outcomes. In the present research we propose a conceptual model that makes salient the separability of positive and negative aspects of relationships during support. This framework suggests that an individual's social support network can be heuristically classified as sources of positivity (i.e., support), negativity (i.e., aversion), or both positivity and negativity (i.e., ambivalence). This framework, combined with existing relationships and health models, makes salient mechanisms linking relationships to health at different levels of analysis. Thus, the major aims of this application are to (a) examine the utility of this model for analyzing the health effects of relationships, and (b) to elucidate basic psychological, behavioral, and physiological mechanisms linking relationships to cardiovascular outcomes. We pursue these aims in five integrative studies. In three studies, we investigate the links between differing relationships and cardiovascular reactivity and recovery in response to acute stress by utilizing social-cognitive paradigms that (a) manipulate perceptions of relationship quality and (b) disrupt or activate psychological processes associated with existing relationships. A fourth study will address the generalizability of these lab-based protocols by testing the association between social relationships and daily life ambulatory blood pressure (ABP) following work in married couples. A final study will examine the links between relationships and recent immune markers of inflammation that confer increased risk for cardiovascular disease in an older, more health-relevant population of married couples. All of the above studies will assess basic mechanisms to help elucidate the processes by which relationships may influence cardiovascular function in each paradigm. Importantly, our preliminary research provides evidence for the feasibility and potential of the proposed research.